122 The Rev. Prof. O'Brien on the Prismatic Spectrum. 



E the commencement of the orange, F the commencement of 

 the yellow, B, C, D all in the red, the tint from A to D being 

 decidedly red, and from D to E verging towards orange. 

 Draw the lines A a, B 6, E C c, D rf, F/ parallel to O X. 

 Then the lines Oa, Oh, O c, O d, Oywill represent the in- 

 dices of refraction corresponding to the different values of « 

 represented by the lines drawn from a, 6, c, r/,yparallel to O X 

 to meet the curve. Hence the spaces ah, he, cd, ^^will re- 

 present the different portions of the prismatic spectrum from 

 the beginning of the red to the beginning of the yellow. From 

 a to c will be decided red ; from c io d will be a mixture of 

 red, reddish-orange and orange, and will appear to be red 

 gradually changing to orange; beyond rf we shall have yellow, 

 and so on. 



38. Now suppose that the absorbing power of the blue glass, 

 through which the spectrum is supposed to be viewed, is such 

 that all the light from B to C, from D to E, and from E to F 

 is completely destroyed, then the appearance of the spectrum 

 will be evidently as follows : from a io h will be red ; from h 

 to c will be a black space of some magnitude ; from c to d will 

 be red of the same tint as at h, since the reddish-orange and 

 the orange, which before were present from c to d, are now 

 obliterated ; from d to^ will be a small black space ; and from 



fio some distance above^ will be pure yellow; and so on. 



Here then we have two reds of very nearly the same tint, 

 a b and c d, separated from eacli other by a broad dark space, 

 he, and the red, cd, separated from the yellow at^by the 

 narrow dark space df. These are exactly the appearances 

 which present themselves when the prismatic spectrum is 

 viewed through a piece of common blue glass of sufficient 

 thickness. 



39. I do not, however, venture to say that the curve A L 

 is precisely of the form here supposed, but only that the ap- 

 pearances we have been attempting to explain are possibly 

 caused by some such peculiarity in the form of the curve, 

 which produces a mixture of colours in some parts of the 

 spectrum. I may observe that the portion cd of the spectrum 

 contains none of the principal fixed lines, and, if I may judge 

 from plates of the spectrum (for I have not at present appa- 

 ratus sufficiently good to exhibit the spectrum with sufficient 

 purity), none but very faint lines occur in this space. I may 

 also remark that the intensity of the light in this space is 

 greater than at any other part of the spectrum. Both these 

 circumstances, as far as they go, confirm the above explana- 

 tion. 



If the spectrum produced by means of a grating before the 



